Accelerating RNA Structure Prediction with FPGA
Author Information
Author(s): Xia Fei, Dou Yong, Zhou Xingming, Yang Xuejun, Xu Jiaqing, Zhang Yang
Primary Institution: National University of Defense Technology
Hypothesis
Can FPGA technology significantly improve the performance of the RNAalifold algorithm for RNA secondary structure prediction?
Conclusion
The FPGA implementation of the RNAalifold algorithm achieved a speedup of 12.2 times compared to the traditional software implementation.
Supporting Evidence
- The FPGA implementation achieved a speedup of 12.2 times over the RNAalifold software.
- Power consumption of the FPGA accelerator was only about 1/8 of that of general-purpose microprocessors.
- The study optimized the RNAalifold algorithm by reorganizing computation order and improving data locality.
Takeaway
This study shows that using special computer chips called FPGAs can make predicting RNA structures much faster than regular computers.
Methodology
The study implemented a systolic array structure on FPGA to optimize the RNAalifold algorithm, focusing on data reuse and load balancing.
Limitations
The study's results are based on specific FPGA hardware and may not generalize to all computing environments.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website